Martin Adámek
Martin Adámek
> Also, you technically need to roll back But why? You have the migrations, they contain the queries, why recompute everything instead of just concatenating?
I'd go with concatenation, as it will be safer (no chance of data loss) as well as faster (no queries involved), and you don't need to touch the snapshot file...
Migrations can contain code that directly works with the database (e.g. `this.execute()` calls), so I would rather go with direct code manipulation instead. It might be a bit tricky to...
Any reason you don't just use `schema:update --dump`?
Do you think it's safe to do it in a non-major release? I know those are internals, but I can imagine people reading those files manually too. Thoughts @metalwarrior665?
That's because it is a static method, the ToC in right sidebar never prints static symbols. We could introduce instance level method with the same name to get around it.
Because it is basically just a shortcut for reading the `INPUT.json` file from the project root, the instance method would do exactly the same. It's indeed more confusing than it...
Note that you are not awaiting the `crawler.addRequests([next.value])` call, that's another problem in the repro.
> We could also propagate the original error message into the status: Yeah, but that's not really about crawlee, right? Those are coming from the platform. We can surely improve...
> platform: "browser", I guess this itself won't help, but crawlee is a Node.js framework, you can't run it in a browser, so compiling it with platform: 'browser' seems just...